MusicDocs is an ongoing web-based application designed to support a broad spectrum of Optical Music Recognition (OMR) tasks based on the application of state-of-the-art models. MusicDocs is designed for flexibility: it supports model inference both on server-side and client-side, including in-browser execution via Open Neural Network Exchange (ONNX). This hybrid architecture ensures scalability and allows for privacy-preserving local processing when needed. The application is available at: musicdocs.ai